Tag Archives: Spring 2011

Using Evolutionary Algorithms for Numerical Optimization of Spacecraft Trajectories

Thursday April 21, 2011 4:00-5:00 p.m.
ME-EM building, Room 112

Dr. Bruce A. Conway
University of Illinois-Urbana

There has been significant progress in the development of numerical methods for the determination of optimal trajectories for continuous dynamic systems, especially in the last 20 years. In the 1990s the principal contribution was new methods for discretizing the continuous system and converting the optimization problem into a nonlinear programming problem. This has been a successful approach that has yielded optimal trajectories for very sophisticated orbit transfer problems. In the last 10-15 years researchers have applied a qualitatively different approach, using evolutionary algorithms, to solving similar problems. Evolutionary algorithms use the principle of “survival of the fittest” applied to a population of individuals representing candidate solutions for the optimal trajectories. In this paper the advantages and disadvantages of these recently developed methods are described and an attempt is made to answer the question of what is now the best extant method.

Yielding of Solid Foams: An Energy Based Approach

Thursday April 14, 2011 4:00-5:00 p.m.
ME-EM building, Room 112

Dr. Murat Vural
Illinois Institute of Technology

Cellular solids such as solid foams and lattice truss structures draw a growing interest in engineering community mainly because of their lightweight. Reliable prediction of yield behavior in cellular solids is critical to the effective use of this class of highly porous materials in designing lightweight and potentially multifunctional structures in a multitude of engineering applications. Although there exist several yield criteria proposed in the literature for solid foams, they are all phenomenological in nature, rely on relatively long list of model parameters that require difficult experimentation not readily available to end user, and none of them can handle the anisotropy observed in commercially available solid foams. Furthermore, a rigorous validation of proposed models against experimental and/or computational data under complex stress states still remains to be a critical issue.

In this talk, a different approach is taken by hypothesizing that the yielding of stochastic foams is governed by the total elastic strain energy density. Based on an analytical framework developed around this hypothesis we propose a pressure-dependent yield criterion for transversely isotropic solid foams. Besides accommodating anisotropy this new energy-based yield criterion renders an advantage over various phenomenological relations proposed in literature by relying only on the elastic prope11ies and uniaxial yield strengths of the material. The results of an extensive finite element analysis and one-of-a-kind multiaxial experiments will be discussed to validate the proposed model and better understand the nature of pressure-dependency in deformation and yield behavior of solid foam s. It will be further shown that proposed analytical framework allows the introduction of new scalar measures of stress and strain (called characteristic stress and characteristic strain) that are capable of representing the elastic response of anisotropic foams with a single elastic master line under arbitrary multiaxial loading paths.

Broadband Mechanics of Structures Under Shock

Thursday April 7, 2011 4:00-5:00 p.m.
ME-EM building, Room 112

Dr. Jason Foley
Air Force Research Laboratory

The realm of impact and penetration presents a challenging set of environments for the design of shock survivable electromechanical systems. Survivable systems must be designed to withstand this harsh environment; intelligent survivable systems and their test instrumentation must also possess the capability to sense and interpret both internal and external dynamics during such events. This presentation explores the state of the art in shock experimentation, simulation, and analysis of these survivable systems.

The talk will begin with a brief survey of a few of the many outstanding research challenges in the world of shock mechanics. Next, the tools used in penetration and impact research will be discussed, with an emphasis on shock test methods and high bandwidth instrumentation (sensors and data acquisition). Results from some structural dynamics experiments are also discussed, with an emphasis on the role of interfaces in the nonlinear response of the systems under study. Two reduced-order computational approaches, wavelet- and Fourier-spectral element modeling, are presented that promise to increase the fidelity of impact/penetration simulations. Finally, new applications in broadband structural dynamics such as microsecond structural health monitoring and wavelet-based system analysis, will be examined. Gratuitous explosions are shown throughout the presentation.

Challenges in Developing Software for Space Trajectory Optimization Problems. Closing the Gap between Theory and Practice.

Thursday March 31, 2011 4:00-5:00 p.m.
ME-EM building, Room 112

Dr. Juan Senen Senent
Research Engineer at Odyssey Space Research

Consider designing a spacecraft trajectory that will visit as many asteroids as possible using the least amount of propellant, in the shortest amount of time, with only a tiny engine. Now, consider designing the trajectory of a spacecraft that collects as many pieces of space debris using minimum propellant. What these problems have in common is that in order to solve them, a large and complex optimization problem has to be solved. Developing software to solve these problems in a fast and accurate way is in itself a challenge. But if we take into account that the software has to be general enough to solve other trajectory optimization problems, should incorporate state of art algorithms while maintaining legacy code, should run in single computing environments and also in computer clusters and should also help the designer to understand the solution by graphically interacting with it, the software development process becomes as complicated as the problem it is trying to solve. This presentation will outline some of the challenges encountered while developing the Copernicus trajectory tool as well as examples of trajectory optimization problems solved with this tool.

Erythrocyte Polarizability: Nonlinear Electrokinetics in Medical Microdevices

Thursday March 24, 2011 4:00-5:00 p.m.
ME-EM building, Room 112

Associate Professor Adrienne R. Minerick
Chemical Engineering-Michigan Technological University

The controlled manipulation of bioparticles with electric fields has become ubiquitous in biotechnology laboratories, usually in the form of gel or capillary electrophoresis. However, creative use of electric fields for bioparticle manipulation in microdevices has also opened doors to opportunities in medical diagnostics. Such analytical microdevices are also referred to as lab-on-a-chip or micro total analytical systems (uTAS) and offer the potential for rapid, point-of-care diagnostics with implications in geographically remote medical situations. With relatively small electric fields, neutral matter can be manipulated: cells can be accelerated or trapped, proteins can be separated, DNA can be precisely focused, and other separations/purifications can be conducted at low concentrations.

This talk wilI focus on a subset of electrokinetics known as dielectrophoresis (DEP) for the analyis of erythrocytes (red blood cells). Dielectrophoresis is the polarizability of neutral & charged particles in a non-uniform alternating current electric field or a spatially non-uniform direct current field. Recent work in our lab has demonstrated that DEP can be utilized for rapid blood typing, recognition of membrane molecule expression, and hemolysis. Two microdevices have been tested with the ABO-Rh blood groups including a batch AC perpendicular electrode confrguration and a continuous sorting DC microchannel with bifurcation and remote electrode configuration. Membrane molecule expression has been systematically tested by comparing the DEP signatures of neat erythrocytes and those subjected to a B(1 -3)-galactosidase enzymatic reaction to digest the A and B polysaccharides from the membrane surface. The presence or absence of the Rh transmembrane protein is also discernable from the DEP signatures. Lastly, hemolysis has also been demonstrated in the low frequency range. By merging the fields of microfluidics, electrokinetics, and biology, handheld microdevices capable of rapidly screening and quantifying diseases, infections, or other ailments may one day become commonplace in medicine.

Nanoscale Thin Film Evaporation and Water Transport for Efficient Energy Systems

Thursday March 17, 2011 4:00-5:00 p.m.
ME-EM building, Room 112

Dr. Shalabh Maroo
Massachusetts Institute of Technology

Novel cooling technologies are required which can meet present cooling demands for concentrated solar irradiation levels of over 1000 sun in concentrated photovoltaic systems to enable major advancements in solar conversion technologies as well as many other important energy systems. Nucleate boiling heat transfer and thermal management devices based on micro-thin film evaporation have shown heat fluxes of only ~ 5 MW/m2. My research focuses on utilizing the concepts of disjoining and capillary pressures in nano-thin film evaporation to achieve ultra high heat fluxes. Nano-film evaporation occurs at the interline region in the three-phase contact line during bubble nucleation and growth. Molecular dynamics simulations were performed to study the evaporation of a nanoscale meniscus. Non-evaporating thin film formation, Hamaker constant of the film and ultra high heat fluxes(~ 100 MW/m2 ) were obtained, and the existence of high absolute negative liquid pressure in nano-films was confirmed. This property of sustaining high negative pressures in liquids at nanoscale can be engineered to provide passive transport of liquid, and applied in power devices to attain significantly higher heat rejection rates.

Transport of water in nanometer and sub-nanometer pores is of fundamental interest. Zeolites, which are porous aluminosilicate minerals, provide a perfect crystalline structure towards this research goal. The aim is to achieve desalination by size-based exclusion of solvated ions through zeolite created membranes, which has the potential to advance desalination technologies. MFI zeolites, which have a pore diameter of about 0.56 nm, are being studied via molecular dynamics simulations and experiments. As the size of a water molecule is roughly 0.28-0.3 nm in diameter, the pore size restricts the flow of water into a molecular chain presenting an extraordinary physical behavior. The water molecular structure inside these subnanometer pores is determined, and the effect of surface charges on water transport is evaluated based on energy analysis. It is also shown that very local energy equilibrium of one water molecule can affect the overall transport behavior in the pores.

Agent-Based Informatics for Autonomous Microgrids

Thursday March 3, 2011 4:00-5:00 p.m.
ME-EM building, Room 112

Dr. Steven Y. Goldsmith
Distinguished Member of the Technical Staff at Sandia National Laboratories

It is a foregone conclusion among power systems experts that localized microgrids based on fine-grained distribution of small capacity generators and fine-grained control of loads using advanced power grid informatics will limit the scope of cascading failures, reduce transmission losses and C02 emissions, enable renewable source penetration, and generally improve the reliability of electrical service throughout the entire grid. However, the current centralized grid coordination and control architecture is inadequate for managing distributed power grids. Centralized SCADA functions requiring substantial human-in-the-loop decision and intervention cannot scale, i.e. cannot decide proper actions, nor decide in time, for the number of independent generation and load resources implied by highly distributed autonomous microgrids. This talk will present an overview Sandia’s technical approach to the power grid informatics challenges posed by distributed autonomous microgrids. The basis of our approach is to leverage advanced multi-agent system (MAS) technology to develop the collective decision functions and interaction protocols necessary to harmonize the competing and common interests among, and coordinate the actions of, a heterogeneous society of autonomous power agents. The presentation will exhibit the general MAS architecture and highlight some fundamental research issues being investigated through development testing on a small computational cluster.

Consumer Needs Engineering: Incorporating Perceptual Aspects of Human Assessment in Product Decisions

Thursday February 24, 2011 4:00-5:00 p.m.
ME-EM building, Room 112

Dr. John Cafeo
GM -Vehicle Development Research Lab

A critical need in designing a consumer product, like an automobile, is the use of credible
quantitative models early in the vehicle design process that quantify the risk of consumer
acceptance given a particular vehicle design. In order to quantify the risk, it is necessary
to be able to predict consumer assessments of a vehicle design as well as estimate the
probability a consumer will assess the design in the same way as the prediction. By
aggregating the risk, the vehicle development teams can logically and defensibly balance,
or tradeoff, the various design attributes and features in order to target a vehicle at the
desired population of customers.

In this talk, I will define this area of work and discuss the challenges in modeling a
perception of an attribute of the vehicle. An intertwined concept is human decision
making. I will discuss the large framework for decisions to help motivate the work. The
development of a model will necessary include a description of the product as well as a
description of the people. Currently, data for this type of model comes from consumer
clinics and surveys. Augmenting this data with anthropometric data as well as
physiological data may allow the estimation of a better model. Because gathering this
type of data is expensive, it is important to be able to use data from multiple events. I will
show an example of a model for head clearance to show one way this might be done.

Volumetric 3-component Velocimetry Measurements near a Rushton Turbine in a Stirred Tank Reactor

Thursday February 17, 2011 4:00-5:00 p.m.
ME-EM building, Room 112

Associate Professor Kendra Sharp
Oregon State University

Prof. Sharp’s presentation will focus on three-dimensional measurements of the flowfield in a stirred tank reactor, however, several other current projects will also be briefly introduced. These projects include: induced-charge electroosmosis for flow control in microchannels; a microchannel-based dialyzer; high-Reynolds number flow in a 90° bend; and ‘ micro’-hydro power for rural electrification.

Volumetric 3-component velocimetry measurements have been taken of the flow field near a Rushton turbine in a stirred tank reactor. This particular flow field is highly unsteady and three-dimensional, and is characterized by a strong radial jet, large tank-scale ring vortices, and small-scale blade tip vortices. The experimental technique uses a single camera head with three apertures to obtain approximately 15,000 three-dimensional vectors in a cubic volume. These velocity data offer the most comprehensive view to date of this flow field, especially since they are acquired at three Reynolds numbers. Mean velocity fields and turbulent kinetic energy quantities are calculated. The volumetric nature of the data enable tip vortex identification, vortex trajectory analysis, and calculation of vortex strength. Three identification methods for the vortices are compared. The visualization of tip vortices up to 140 degrees past blade passage in the highest Reynolds number case is notable and has not previously been shown.

A Laboratory for Dynamic Systems and Human-Robot Interface

Thursday February 3, 2011 4:00-5:00 p.m.
ME-EM building, Room 112

Assistant Professor Mohammad Rastgaar
Michigan Technological University

In this talk, I will introduce two areas of my research at Michigan Tech. In the first part, “Orthogonal Eigenstructure Control for Active Vibration Cancellation”, I will present Orthogonal Eigenstructure Control (OEC) and its application to structural vibration cancellation. OEC is a feedback control method applicable to general multi-input multi-output linear systems. While the available control design methodologies for large flexible structures offer a large and complex design space of options, this control methodology offers a significant simplification of the design task while still allowing some experience-based design freedom. Specifically, OEC eliminates the need for shaping and defining the desirable eigenstructure as required in eigenstructure assignment methods. The mathematical aspects of orthogonal eigenstructure control as well as the results of implementing this method to cancellation of vibrations in a steel test plate under tonal and wideband disturbances will be presented. In the second part of the talk, “Stochastic Estimation of Multi-variable Human Ankle Mechanical Impedance”, I will present a lower extremity rehabilitation robot and the results of estimation of multi-variable dynamic ankle impedance in humans with different types of muscle activities utilizing this robot. The results suggest that different dynamics may govern standing and walking which needs to be considered in the design of lower-extremity assistive or prosthetic robots.